NEW YORK – A group led by researchers from the Institute for Systems Biology in Seattle has developed a metabolic method that identifies circulating tumor cells (CTCs) in lung cancer patients by measuring the level of a biomarker called hexokinase-2 (HK2).
The team believes the method could eventually be used to help establish cancer patient prognosis before determining the right treatment.
Current methods to sort and detect CTCs in a patient's liquid sample either rely on physical properties, such as Angle's Parsortix/HyCEAD platform, or epithelial biomarkers like Menarini Silicon Biosystem's CellSearch instrument. The methods have been successful at detecting CTCs derived from tumors such as prostate and breast cancer.
However, conventional methods do not often detect CTCs derived from lung cancer, said Wei Wei, associate professor at the ISB and a developer of the new metabolic marker method. He noted that physical-based methods are not specific enough, so "you might get other compounding cell types, like circulating endothelial or mesothelial cells."
His team instead chose to monitor certain biomarkers linked to increased glucose consumption in CTCs, as the cells tend to boost glucose metabolism to fuel their uncontrollable growth. While most normal cells use hexokinase-1 for glycolysis, many tumors express elevated levels of HK2.
"Hexokinase-2 is the first enzyme that converts glucose to glucose to glucose-6-phospate, or the first step of glucose metabolism," Wei said. "We found it interesting because [HK2] has a restrictive distribution in normal tissue but [is] upregulated in cancer cells."
Wei's team therefore exploited the presence of HK2 as a surrogate for metabolic activity in CTCs by building a workflow to measure the amount of lung cancer-derived CTCs in a patient's bloodstream.
The method starts with enriching and centrifuging a patient's blood sample to remove red blood cells, platelets, and CD45-positive cells. Wei's team recovers enriched cells and places them on a polydimethylsiloxane (PDMS) microwell chip that the group built for cell fixation, permeabilization, and immunostaining of several biomarkers including HK-2.
Wei's team then washes and images cells on the chips, identifying CTCs based on their fluorescence intensity. He said that each chip contains about 10,000 microwells that can be individually scanned and imaged at single-cell resolution within 15 minutes.
Targeted CTCs are then individually retrieved by a motorized micromanipulator based on their recorded address for single-cell sequencing to validate their malignancy.
Prior to sequencing, Wei said the workflow can be performed within two to three hours. He noted the workflow needs to be "relatively quick" due to the metabolic process and degradation of CTCs.
The method uses about 5 ml of a patient's blood sample, which Wei believes is enough to gather a "reasonable number of CTCs" per patient." The single-cell sequencing portion then requires between one to two weeks to complete.
"We tried to push the limit of detection to see if we can detect CTCs from a small volume of blood because we think it should be more sensitive than prior methods," Wei said. "Of course, you could use 10 ml or higher, but the smaller the amount, the more valuable it is [in the clinical space]."
In a proof-of-concept study published earlier this month in PNAS, Wei and his colleagues investigated the use of single-cell HK2 detection as a marker to identify CTCs in peripheral blood and other liquid biopsy samples from lung adenocarcinoma (LUAD) patients.
After confirming the metabolic method's applicability through blood samples spiked with LUAD cell lines known to express levels of HK2, the group used the method to identify CTCs in blood samples from a cohort of 50 treatment-naïve stage III-IV LUAD patients. They detected CTCs in 72 percent of the patients (with five or more CTCs detected in 5-ml blood samples in 44 percent of the patients), but detected no CTCs in blood samples from 30 healthy donors.
For patients with positive CTC counts, the team found HK2-high, CK-negative CTCs in 81 percent of the patients, and that the phenotype was the most common in 47 percent of the patients. Wei said his team therefore aimed to see if they could spot CK-negative tumor cells in other samples, including pleural fusions and cerebral spinal fluid (CSF).
Selecting CTC-positive individuals from the lung cancer cohort, the researchers analyzed 10-ml malignant pleural effusion (MPE) samples and several 1.5-ml CSF samples. Both sample types exhibited consistent copy number variant profiles at the single-cell range, but the CK-negative, HK2-high phenotype was a minority population in the patient cohort.
"The difference in CK subpopulations across different types of liquid biopsy may be the consequences of different microenvironments of these liquid biopsy samples [and/or] distinct mechanisms of dissemination between CK-positive and CK-negative cells," the study authors said.
To check whether CK-negative CTCs acquired mesenchymal features and become more invasive, Wei's team performed single-cell RNA sequencing of CK-positive and CK-negative CTCs in a MPE sample. However, they found that the CTC types could not be respectively linked to epithelial and mesenchymal phenotypes.
Wei and his colleague then decided to analyze the differential transcriptomic profiles between CK-outlier cells in epithelial and mesenchymal tumor cell populations. They saw that CK-low cells exhibit metastasis and EGFR-TKI resistance-related molecular signatures compared with CK-high cells, regardless of their epithelial-mesenchymal transition status.
Encouraged by the transcriptome signatures of the CK-low cells, Wei's team explored the difference in clinical outcomes between LUAD patients with different numbers and ratios of HK2-high/CK-negative CTCs. Classifying 36 of the HK2-high/CK-negative LUAD patients into cohorts based on their CTC percentages, the team found that the number of metastatic sites was higher for HK2-high/CK-negative than HK2-high/CK-positive CTCs.
The results also suggested that patients with CK-negative CTCs had a shorter progression-free survival in patients compared to their CK-positive counterparts, which Wei said might explain the poor therapy responses of certain patients that have been reported in prior studies.
"This also suggests that those CTCs tend to transition to CK-negative to facilitate the process," Wei added. "Patients with EGFR-L858R driver oncogene mutations are more likely to have CK-negative CTCs in the blood, compared to patients with other mutations, like EGFR-19-Del."
Wei's team plans to further explore HK2's utility as a prognostic marker in larger lung cancer cohorts to predict whether a patient has a good therapy response. The group will analyze the tumor biology by interrogating the underlying transition from CK-positive to CK-negative CTCs and the associations between the different EGFR phenotypes.
"If we can understand how the CTCs are triggered to the epithelial-mesenchymal transition, it might help us determine whether CK-19 could be interrogated by drugs," Wei said. "If we can find the master regulator for this transition, we can stop the use of downstream effectors to stop the transition and make it less metastatic."
Wei said that his team will also see if CK-negative CTCs are indicators of poor prognosis or metastasis, or hold drug-resistant features in other cancer types. He noted that his team has preliminary data that CK-negative CTCs can exist in other tumor types — including non-small cell lung cancer and bladder cancer — and has launched projects that will validate the metabolic method in these cancers.
"We can get exfoliated tumor cells from urine samples as a diagnostic marker for bladder cancer, [but] are still looking at HK2 behavior and other CK-negative CTCs in urine samples," Wei said.
Shana Kelley, a biochemistry professor at the University of Toronto who was not involved with the study, agreed that CTC levels in lung cancer patients are much lower than in other cancer types. She believes that the HK2 metabolic marker could potentially make lung cancer detection much more accurate and allow a new approach for phenotypic liquid biopsy.
Kelley and her colleagues at the University of Toronto have launch a startup called Cellular Analytics to capture CTCs and rare immune cells to establish drug sensitivity for oncology drug development, starting with lung cancer.
Wei's team has filed a patent for the HK2-based metabolic method through the Institute for Systems Biology. He said the group is potentially interested in commercializing the method following the completion of larger validation studies in the future.